diff --git a/.gitmodules b/.gitmodules
index d5788ea..5a78519 100644
--- a/.gitmodules
+++ b/.gitmodules
@@ -1,3 +1,3 @@
 [submodule "ggml"]
     path = ggml
-	url = https://github.com/leejet/ggml.git
+	url = https://github.com/ggml-org/ggml.git
diff --git a/ggml b/ggml
index 7026887..5fdc78f 160000
--- a/ggml
+++ b/ggml
@@ -1 +1 @@
-Subproject commit 70268874234fa06cf058739a221bfbe94129f330
+Subproject commit 5fdc78fff274094e2a1b155928131983362d8a71
diff --git a/model.cpp b/model.cpp
index 1febed9..575cb3a 100644
--- a/model.cpp
+++ b/model.cpp
@@ -1055,21 +1055,7 @@ bool ModelLoader::init_from_gguf_file(const std::string& file_path, const std::s
     gguf_context* ctx_gguf_ = NULL;
     ggml_context* ctx_meta_ = NULL;
 
-    auto on_tensor_shape_read = [](const int64_t* ne, uint32_t n_dims, struct gguf_tensor_shape* shape) -> bool {
-        for (int i = 0; i < GGML_MAX_DIMS; i++) {
-            if (i < n_dims) {
-                shape->ne[i] = ne[i];
-            } else {
-                shape->ne[i] = 1;
-            }
-        }
-        for (int i = GGML_MAX_DIMS; i < n_dims; i++) {
-            shape->ne[GGML_MAX_DIMS - 1] *= ne[i];  // stack to last dim;
-        }
-        return true;
-    };
-
-    ctx_gguf_ = gguf_init_from_file_ext(file_path.c_str(), {true, &ctx_meta_}, on_tensor_shape_read);
+    ctx_gguf_ = gguf_init_from_file(file_path.c_str(), {true, &ctx_meta_});
     if (!ctx_gguf_) {
         LOG_ERROR("failed to open '%s'", file_path.c_str());
         return false;